Meta is rolling out new safety features designed to give teenagers more age-appropriate experiences when using its AI chatbot, with a stronger focus on protecting young users during sensitive conversations.
One of the biggest updates allows Meta AI to notify parents if a teenager's conversation suggests they may be at risk of self-harm or suicide.
Previously, the chatbot would encourage teens to contact a trusted adult or reach out to crisis helplines. Under the new system, Meta AI can also alert a supervising parent when it detects signs of emotional distress, even if the references are subtle.
According to Meta, the feature was developed with input from parents, safety experts and mental health professionals to better identify conversations that may require additional support.
To reduce the risk of false alarms, every conversation flagged by the AI will be reviewed by a human before a notification is sent to parents.
If there is still uncertainty after the review, Meta said it will prioritise safety by sending the alert rather than risking a missed opportunity to intervene.
The company is also developing a system that could contact emergency services if any user, including teenagers or adults, appears to be at imminent risk of suicide during conversations with Meta AI.
Meta said more than 75 mental health clinicians specialising in adolescent psychology helped review hundreds of AI responses to ensure the chatbot responds with empathy while guiding users towards appropriate offline support.
Alongside these updates, Meta is introducing stricter content controls for teenagers. While teen accounts already have restrictions on conversations involving sexual or romantic topics, parents can now choose a"Limited Content" setting that further reduces the types of prompts Meta AI will respond to.
The latest features reflect Meta's ongoing efforts to create safer AI experiences for younger users while giving parents greater visibility when serious concerns arise.